I don’t know how dirty it is, or, indeed, how secret, but one key point that one must be very careful not to fool oneself on when working on a Linked Data project (like FishDelish) is the relative contribution of
- using “linked data principles”
- using RDF
- using a standard format
- the fact that Web apps are more sophisticated yet easier to build than ever before
- that attention and effort begets improvements.
For example, in my last post, I pointed out that the Fish On Line book isn’t really very modern in its interaction mechanisms (nor, frankly, in it’s design). We plan to rectify that and one step toward that rectification is migrating FishBase data to RDF and serving it up with a SPARQL endpoint. But, pretty clearly, that migration is not a necessary step to improving the layout, or even improving the interaction mechanisms! It’s not even a necessary step to “linking” the data (either in the sense of producing data mashups or in the sense of providing auxiliary connections).
We don’t have clear evidence (nor will is it a direct aim of this project to provide such evidence) that it is overall easier or more productive or more effective to improve Fish On Line via standard Linked Data techniques! I do know it’s probably easier for me, especially now as there’s a growing set of reasonable tools. I also don’t know if we could have gotten funding to do it some other way (certainly not from this JISC program).
Is this a big problem? It depends on what you want to know or do. I think it’s likely to be helpful to FishBase (which is good). I think there’s some clear wins from the approach we’re taking (also good). It’s pretty fun (hey, that’s a good!). There’s a good chance of pushing species data curation and integration forward (very, very good). I think we can get some active essay pedagogic joy (hurray!).
And maybe, just maybe, we can get a reasonable case study out of it. But case studies are tricky research objects, fraught with subjectivity and possible sources of bias. Generalising is also difficult. Such research is really hard work and somewhat orthogonal to other goals. However, given the strong interest in Linked Data going on at the moment (esp. given the Open Data movement…which is not quite the same), there seem to be good opportunities to do some of that research. FishDelish is, I think, a pretty good case to try to analyse.
This post might seem a bit grumpy, but it really isn’t! I just want to be clear as possible about what conclusions, with what degree of certainly, we can reasonably draw from various events. Applying a skeptical eye to one’s own work, claims, efforts is an important step in gaining that clarity.